The metastasizing and hormone response properties of breast cancers largely determine survival of patients with the disease. Our objective is to establish the origins and lineages of malignant and hormone unresponsive tumor cells starting with the normal cell that is transformed. Our proposal builds on the idea that it is the phenotype of the normal cell that is transformed which determines the malignant and hormone response phenotypes of resulting neoplastic cells and not the random variations that may occur subsequently in tumors. This distinction between deterministic and stochastic models of neoplastic progression is fundamental to theestablishment of thoroughly effective strategies for cancer control. We propose to test the hypothesis that hormone "unresponsive" tumor cells are not unresponsive at all; they derive from the transformation of a corresponding non-estradiol response phenotype in normal mammae, not from random adaptation and selection of estradiol response variants during tumor growth. We also propose to test the hypothesis that the degree of commitment of a normal cell to terminal secretory differentiation (degree of variogenicity) at transformation determines the malignancy of the neoplastic derivative(s). This study is made feasible by establishing in culture a clonal lineage of normal mammary cells (found in murine and human mammae) which we call the E-G-D lineage. This lineage is composed of multiple hormone response phenotypes of which an estradiol responsive phenotype is a progenitor. A terminal differentiation sequellae can be induced by glucocorticoids via a Ca++ sensitive pathway. By uncoupling terminal differentiation from growth we are able to cause normal cells to neoplastically transform in vitro. (C)